►
From YouTube: NuPIC Sprint Planning - Aug 22, 2014
Description
Where we plan Sprint 29: http://status.numenta.org/issues.html#milestone=Sprint+29
A
Hello
new
pick:
this
is
sprint
planning
session
for
sprint
29,
it's
august,
22nd
2014
and
let's
jump
right
in
get
my
screen
shared.
A
So
these
are
the
closed,
pull
requests
from
sprint
28,
which
is
a
sprint
we're
closing
out
right.
Now,
I'm
going
to
go
over
some
of
the
main
essential
repos.
Just
a
few
of
these,
like
nuke
core,
for
example,
we
added
back
a
compile
flag
that
shows
errors
on
unused
function.
Parameters
mark
did
that
austin
added
a
sequels
plus
only
region
example
and
utensil
fixed
a
build.
B
A
I
removed
a
class
that
was
never
loaded
in
the
swarming
code
and
I
also
removed
parallel
unit
test
runs,
which
happened
to
also
swallow
syntax
errors,
so
we'll
try
and
get
those
that
re-enabled
at
some
point,
because
it
did
save
a
few
minutes
of
time
on
the
build.
A
Jayden
did
some
cleanup
with
renaming
and
temporal
memory
and
added
errors
to
the
hot
gym
example.
Some
of
the
other
things
if
you've
noticed
we've
got
a
new
repo,
which
is
the
java
port.
So
there's
I've
been
working
with
david
ray
who's,
the
the
original
author
of
the
java
port
to
get
that
that
repository
into
a
good
state
for
reporting
on
all
and
testing
and
building
and
all
that
stuff,
so
a
little
bit
more
work
to
do
there,
but
mostly
it's
just
david
ray
continuing
to
work
on
algorithm
implementation.
There.
A
The
other
stuff,
is
just
kind
of
tooling
improvements.
Content
updates
on
nementa.org,
nothing
extremely
important.
So
those
are
the
closed
issues.
Here
are
the
open
pr's
that
are
currently
in
the
sprint,
so
I
want
to
do
a
quick
review
of
those.
Most
of
these
are
still
just
still
under
work
and
will
be
moved
to
the
next
sprint,
like
the
stuff
that
jim
bridgewater
is
working
on.
Mark
mark
is
continuing,
he's
kind
of
back
from
sabbatical
and
continuing
on
the
anomaly
modular
pr.
A
B
A
R
pr's
about
the
let's
see
stripped
unlearned
columns.
I
think
these
this
is
for
core
and
there's
also
one
for
new
pick
as
well,
so
those
are
still
under
work.
I
think
this
is
green.
A
I
don't
know
if
you're
ready
to
merge
that
scott,
but
it
looks
like
it's
good,
very
close,
very
close,
okay
and
then
there's
an
associated
one
one
of
these
bottom
ones.
That
is
the
new
complement
to
that.
A
We
had
a
short
discussion
on
the
mailing
list
about
a
default
configuration
value
for
the
sp's
wraparound
parameter,
so
jim
bridgewater
is
forging
ahead
with
that
change.
That's
what
these
are
utensil
and
another
guy.
I
think
his
name
is
richard
or
robert
crowder,
richard
crowder
a
little
bit
more
attention
for
the
windows
support,
so
they
are
both
sort
of
now
and
then
actively
contributing
to
that
that
build
is
passing
and
they've
got
an
app
conveyor,
build
that's
trying
to
run
on
windows.
A
So
that's
an
ongoing
nipit
core
enhancement
for
the
build.
It's
got
some
test
refactoring
and
there's
that
other
one
I
mentioned.
That's
the
component
complement
to
that
one,
and
I
haven't
heard
back
from
utensil
about
the
c
plus
plus
tests.
I
did
send
him
an
email
asking
him
how
he
thinks
we
should
move
ahead
with
that,
so
that's
still
kind
of
on
hold,
so
those
are
the
currently
open
pr's.
I'm.
A
I
think
all
of
these
I'm
going
to
push
ahead,
see
pull
requests
open
and
we
will
bump
all
these
in
newpick
to
spread
29.
A
Okay,
now
we're
going
to
look
at
the
currently
open
issues
in
spring
28.
I've
done
some
cleanup
beforehand,
like
I
always
tried
to
do,
or
else
this
meeting
would
last
just
too
long.
Here's
one
last
one.
It
needs
to
split
29.
A
So
open
issues
in
sprint,
28,
let's
go
buy
repo
here,
nupit
core
the
wrap
around
parameter.
Like
I
mentioned,
that's
the
issue
for
that
now
this
spatial
pooler
behavior
deviates
from
the
python
implementation.
I
do
believe
he
has
a
pr
for
that
and
it's
a
test.
I
think.
No,
that's
not
it.
There
was.
There
was
one
pr
scott
that
I
that
I
commented
on
that
jim
added.
A
I
think
it
was
on
in
new
pic,
but
jim
added
a
test
that
failed
and
I
think
it
seemed
to
be
a
legitimate
test
but
yeah
the
sp
permanent
test,
and
I
want
to
know
what
is
your
take
on
moving
forward
with
this,
because
I
think
the
test
is
valid
and
it
and
it
uncovers
a
condition
that
we
need
to
fix.
Is
that
correct.
C
Yeah,
I
have
another
task
that
I
was
kind
of
using
for
this
and
that
another
issue
that
he
found
but
yeah
definitely
not
a
value
too.
A
Okay,
it's
kind
of
hard
to
hear
you
there
did
was
this.
Is
this
something
that
I
mean?
We
can't
merge
this
while
it's
failing
yeah
yeah.
C
I
think
I
think
it's
definitely
about
the
issue.
I
haven't
looked
at
his
test
specifically,
so
I
don't
know
I'd
have
to
look
at
his
test,
but
yeah
definitely
something
we
should
be
looking
into.
A
C
Yeah,
okay,
yeah!
Is
there
a
is
there
a
issue
for
that.
A
C
Okay,
I
don't
know
if
he's
planning
on
assigning
him
currently,
so
I
don't
know
if
he's
planning
on
I'm
fixing
it
or
not,.
A
Okay,
I'll
I'll
ask
him.
A
A
Okay,
all
right,
we
still
have
c
plus
plus
unit
test
ticket
open.
I
think
I
believe
they
are
generating
x
unit
style,
but
this
is
sort
of
this
ticket
isn't
very
well
defined.
I
might
go
back
and
clean
it
up.
A
A
A
I
thought
this
was
ready
to
go.
No,
no.
This
is
just
an
issue
there,
but
there
is
a
pr,
that's
green
for
it.
I
believe
yeah.
This
is
the
one
I
was
talking
about
earlier.
So
sorry,
you
said:
there's
a
little
bit
more
work
to
do
on
that.
A
Okay,
so
that
those
are
the
new
core
tickets.
These
are
the
new
pick
tickets.
This
top
one
problem
running
python
test
has
become
just
like
a
collection
area
for
lots
of
people's
errors.
A
So
I
at
this
point
I
need
to
break
it
up
into
different
tickets
and
try
and
figure
out
what
the
individual
issues
people
are
having
are.
So
that's
on
me,
some
more
spatial
pooling
stuff
that
jim
opened
language,
independent,
checkpoint
and
serialization.
So
yeah,
quick,
quick
discussion
on
this
I
sort
of
asked
is
this:
was
this
for
an
investigation
and
if
the
investigation
is
done,
should
we
close
it
and
if
so,
what?
What
is
our
next
action.
C
Done
I
was
kind
of
hoping
we
have
like
a
solid
plan
for
moving
forward.
I
I
guess
we
can
close
this
and
I
don't.
I
don't
really
know
how
to
go
about
putting
consensus
on.
A
It's
definitely
past
the
time
so
yeah
it's
it's.
I
just
need
to
close
it,
so
nobody
said
minus
one
on
it:
yeah
yeah!
So
here
I
mentioned
it's
been
over
three
days
now
and
I'm
asking
what
are
our
next
actions
for
this
right,
so
I
think
we
can
assume
that
we're
going
to
move
to
c
plus
plus
11.,
given
that
assumption,
I
think
we
should
close
this
ticket
and
create
a
ticket
for
serialization
using
the
best
option
for
c
plus
11.,
okay,
okay,
so
I'm
I'm
gonna
go
ahead
and
close.
It.
A
I'm
even
gonna
I'll,
just
I'll,
create
a
ticket
after
this
about
the
implementing
the
serialization
method.
Let's
do
this,
let's
just
do
it
right
now
and
we're
gonna
do
this
in
python
right!
Well,
I
ain't
looking
for.
A
So
so
so
we're
going
to
do
the
serialization
in
in
nuclear.
C
There's
different
benefits
for
those,
so
I
can
see
either
way,
but
it
would
be
great
if
we
can
get
out
of
picnic
well
yeah.
Also
there
we
have
seen
some
random
deserialization
bugs
in
the
c
plus
plus,
but
they
are
seem
very
rare
and
you
have
to
be
able
to
reproduce.
C
On
the
python
side,
it
has
to
give
it
a
pickle
and
also
you
can
start
bringing
in
files
with
that
yeah
yeah,
so
they're,
just
different
okay
on
both
sides.
A
C
One
of
the
questions
is
we're
going
to
have
one
set
of
specifications
for
the
format,
and
so
I
assume
that
will
live
in
core
right.
A
I'm
just
I'm
leaving
this
rather
broad
right.
Now
I
don't
know.
Is
this
something
that
we
need
to
do
quickly
or
not?
Is
this
gonna?
Are
we
gonna
be
able
to
work
on
this
in
the
next
sprint,
or
should
I
put
it
into
one
after.
C
I
think
we
should
try
to
decide
on
the
format
that
we're
going
to
use
and
get
you
know,
run
it
by
the
community.
Everyone
know
our
intents
soon
in
the
next
couple
weeks,.
C
A
A
So
we're
getting
some
good
documentation
from
some
people
in
the
community.
So
that's
that's
good!
This.
This
guy
sylvain
is
working
on
some
docks.
He
said
he's
going
to
work
on
cla
model,
so
that's
great
jeff
you're
still
on
tap
to
improve
the
wiki
dock
for
encoders,
which
jayden
did
a
bit
of
so
you
might
not
have
a
lot
of
work
to
do
all
right,
renaming
newpick
to
new
python.
A
I'm
gonna
move
that
ahead
to
the
next,
the
next
next
sprint,
the
other
ones,
though
I
think
we're
just
going
to
move
ahead
unless
someone
sees
something
here
that
they
just
know
is
not
going
to
happen.
E
You
know:
is
there
any
action
from
that
email
thread
that
we
need
to
do.
I
I'm
totally
missing
your
your
point.
What
what
email
thread
are
we
talking
about
the
email
thread.
E
Updating
the
white
paper
and
all
of
that
stuff
cool,
oh
yeah,
yeah,
okay,
is
there
any
action
that
has
to
come
out
of
that.
A
Most
most
likely,
but
I
don't
think
we've
quite
decided
what
we're
going
to
do.
I
haven't
seen
any
responses
from
the
community
since
your
last
email,
I
was
hoping
someone
would
jump
in
and
say
sure,
here's
I
started
the
wiki
fax
for
jeff
and
they
would
start
writing
questions
in
there.
So
I'm
hoping
that
that
happens.
If
not,
I
could
I
could
start
an
empty
document
and
kind
of
poke
them
and
try
and
get
them
to
to
fill
that
in
yeah.
Other
ideas-
that's
fine,
too
yeah
yeah.
E
B
A
I
mean
because
jeff
is
the
one
that
needs
to
take
a
look
at
that
and
and
look
at
the
changes
and
because
it's
in
la
tech-
and
it's
it's
just
really
hard
for
him
to
review
any
of
the
changes
like
I
could
create
a
document
for
them,
but
it
wouldn't
show
a
diff.
So
I
don't
know,
I
think,
that's
sort
of
a
lost
cause
that
new
pick
documents.
A
I
had
high
hopes
for
it
at
the
beginning,
but
I
think,
but
we
do
need
to
just
figure
out
a
way
for
for
jeff,
to
get
feedback
or
for
at
least
for
him
to
share
his
his
research
process
a
little
bit
with
the
community
because
they're
they're
they're
hungry
for
it
you
know
yeah.
So
so
I'm
gonna
wait
and
see
what
happens
on
the
mailing
list
there
and
I'm
sure
some
action
will
come
out
of
it
yeah.
A
A
And
the
other,
the
other
stuff
I'll
clean
up
as
we
go,
I
think
there's
a
couple
of
cerebro:
that's
not
cerebral.
A
Oh,
my
page
is
a
bug,
so
one
thing:
that's
that
I
changed
last
spring:
I
had
a
whole
bunch
of
help
wanted
unassigned
help
wanted
tickets
that
I
put
into
the
current
sprint,
I'm
hoping
that
it
might
make
it
easier
for
people
to
find
something
to
work
on,
but
it
didn't
really
work
and
it
just
kind
of
cluttered
the
sprint
up.
A
So
this
time
I
did
not
do
that.
I
left
I
I
removed
a
bunch
of
where
am
I
there
we
go.
I
I
removed
a
whole
bunch
of
those
help
wanted
tickets.
Most
of
them
were
in
like
cerebro,
2
or
tooling,
or
you
know
newpic.org,
or
something
like
that.
So
I
took
all
those
out
because
it
was.
It
was
cluttering
everything
up
so
now
we
just
got
these
four
tickets
and
four
different
projects
and
we'll
bump
those
up
to
sprint
29.
A
A
And
so
I
I
also
have
some
stuff
in
spring
30
that
I
know
I'm
not
going
to
get
to
or
that
we're
not
going
to
get
to
until
after
the
next
sprint.
So
I'm
planning
ahead
a
little
bit,
but
it's
not
super
important
stuff.
A
Let
me
refresh
this
page
real,
quick
and
we'll
see
the
all
of
the
issues
in
sprint.
29.
A
Okay,
so
we've
got
41
issues,
which
is
pretty
pretty
heavy.
In
my
opinion,
I
may
go
through
here.
Some
of
these
were
already
in
like
austin's
out
today
he
told
me
he
was
going
to
work
on
these
three
issues.
Some
of
them
are
still
unassigned.
Some
of
them
are
marked
with
help
wanted
or
unassigned,
and
those
are
issues
that
I
think
I
would
like
to
get
done,
but
I
don't
know
who's
going
to
do
them,
so
those
are
up
for
grabs
if
anybody
wants
to
work
on
them.
A
Okay,
okay,
so
let
me
just
make
sure
yeah,
so
there's
nothing
in
sprint
28.
So
this
is
our
sprint
29
at
this
point,
and
I
think
we've
hit
all
of
these
issues
one
way
or
another.
Most
of
them
don't
need
a
lot
of
discussion.
So,
okay,
anything
else
from
from
anybody
that
you
want
to
talk
about
in
the
sprint
planning
meeting
before
we
move
ahead.
A
All
right
thanks
for
joining
in
and
if
anybody
wants
to
know
how
they
can
help
go
to
nementa.org
or
email
me
at
matt
demento.org,
and
I
will
explain
great
detail
how
you
can
help
build
new
fake
thanks
guys.